MCPcopy Create free account

hub / github.com/OpenBMB/AgentCPM-GUI / functions

Functions568 in github.com/OpenBMB/AgentCPM-GUI

⨍ Functions 568◇ Types & classes 49
Functionis_click
Args: predict_bbox: Tensor of shape [bs, 2], format [x, y] (coordinates of predicted points). target_bbox: Tensor of shape [bs, 4
eval/grounding_eval/code/Intern2.5-VL/evaluate_grounding_text2bbox.py:110
Functionload_image
(episode,image_path,history_list,use_low_instruction)
eval/run_predict_os_gensis.py:278
Functionload_image
(episode,image_path,history_list,use_low_instruction)
eval/run_predict_os_atlas.py:297
Functionload_image
(episode,image_path,history_list,use_low_instruction)
eval/run_predict_ui_tars.py:357
Functionload_image
(episode,image_path,history_list,use_low_instruction)
eval/run_predict_qwen2_5VL.py:394
Functionload_image
(episode, image_path, data_name)
eval/run_predict_minicpm.py:92
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/OS-genesis/bbox2text_eval_osgenesis.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/OS-genesis/fun2bbox_eval_osgenesis.py:23
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/OS-genesis/text2bbox_eval_osgenesis.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/minicpm/text2bbox_eval_minicpm.py:22
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/minicpm/bbox2text_eval_minicpm.py:22
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/minicpm/fun2bbox_eval_minicpm.py:24
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/UI-TARS/bbox2text_eval_uitars.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/UI-TARS/text2bbox_eval_uitars.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/UI-TARS/fun2bbox_eval_uitars.py:23
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/GPT-4o/text2bbox_eval_click.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/GPT-4o/bbox2text_eval.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/GPT-4o/fun2bbox_eval_click.py:23
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/GPT-4o/text2bbox_eval_gpt-4o_with_grounding.py:23
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/GPT-4o/fun2bbox_eval_gpt-4o_with_grounding.py:23
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/OS-Altas/text2bbox_eval_osatlas.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/OS-Altas/fun2bbox_eval_osatlas.py:22
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/OS-Altas/bbox2text_eval_osatlas.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/Aguvis/fun2bbox_eval_aguvis.py:23
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/Aguvis/text2bbox_eval_aguvis.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/Aguvis/bbox2text_eval_aguvis.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/Qwen2.5-VL/fun2bbox_eval_qwen.py:23
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/Qwen2.5-VL/bbox2text_eval_qwen.py:21
Functionload_image_from_path
从指定路径加载图片 :param image_path: 图片文件路径 :return: PIL.Image 对象
eval/grounding_eval/code/Qwen2.5-VL/text2bbox_eval_qwen.py:21
Functionlocal_balance_proc
Balance data and tasks created in local machine, sync with global.
rft/trainer/zmq.py:45
Methodmonitor
检查缓存中是否有超时的任务
rft/trainer/zmq.py:581
Functionmove_to
(device)
eval/run_predict_os_gensis.py:63
Functionmove_to
(device)
eval/run_predict_odyssey.py:64
Functionmove_to
(device)
eval/run_predict_os_atlas.py:68
Functionmove_to
Move the model to the specified device. Args: device (str): The device to move the model to. Returns: None
eval/run_predict_aguvis.py:431
Functionmove_to
(device)
eval/run_predict_ui_tars.py:68
Functionmove_to
(device)
eval/run_predict_qwen2_5VL.py:77
Functionmove_to
(device)
eval/run_predict_minicpm.py:63
Functionno_sync
For FSPD2, disable gradient synchronization for all model parameters.
rft/trainer/utils/__init__.py:46
Functionpad_batch
(batch: BatchTokensType, pad_id: int, seq_length: int)
eval/utils/utils_odyssey/qwen_generation_utils.py:25
Methodprediction_step
Perform an evaluation step on `model` using `inputs`. Subclass and override to inject custom behavior. Args: mo
sft/trainer.py:46
Methodprediction_step
(self, model, inputs, prediction_loss_only, ignore_keys: Optional[list[str]] = None)
rft/trainer/arl.py:872
Methodprepare_inputs_for_generation
( self, input_ids, past_key_values=None, inputs_embeds=None, **kwargs )
eval/utils/utils_odyssey/modeling_qwen.py:984
Functionprocess_data
Process the data to run predict. Args: episode (Dict): one single data. image_path (str): the image_path processed.
eval/run_predict_aguvis.py:452
Functionprocess_step
(args)
eval/utils/convert_output.py:68
Functionprocess_step_data
Process a single step of data, load prediction results and evaluate. Args: step_data (dict): Data of a single step. Returns:
eval/run_eval_agent.py:94
Methodprovider
为工作进程提供数据
rft/trainer/zmq.py:383
Functionreact_check
(completions, solution: list[dict], resolution, bboxs, step_id, **kwargs)
rft/trainer/utils/gui_eval.py:334
Methodreporter
报告队列状态
rft/trainer/zmq.py:408
Methodreprocess
处理任务并生成处理后的数据块
rft/trainer/zmq.py:376
Functionrun_episode
(episode, image_path,history_list,use_low_instruction)
eval/run_predict_os_gensis.py:273
Functionrun_episode
(question, episode)
eval/run_predict_odyssey.py:411
Functionrun_episode
(episode, image_path,history_list,use_low_instruction)
eval/run_predict_os_atlas.py:292
Functionrun_episode
(episode, image_path,history_list,use_low_instruction)
eval/run_predict_ui_tars.py:352
Functionrun_episode
(episode, image_path,history_list,use_low_instruction)
eval/run_predict_qwen2_5VL.py:367
Functionrun_episode
(episode, msg,)
eval/run_predict_minicpm.py:73
Methodsave_vocabulary
Save only the vocabulary of the tokenizer (vocabulary). Returns: `Tuple(str)`: Paths to the files saved.
eval/utils/utils_odyssey/tokenization_qwen.py:239
Methodserve_stealing
处理其他节点的任务窃取请求
rft/trainer/zmq.py:415
Methodset_input_embeddings
(self, new_embeddings)
eval/utils/utils_odyssey/modeling_qwen.py:649
Methodset_output_embeddings
(self, new_embeddings)
eval/utils/utils_odyssey/modeling_qwen.py:981
Methodstream_generator
()
eval/utils/utils_odyssey/modeling_qwen.py:1205
Functionswitch
(val1, val2, boolean)
eval/utils/utils_odyssey/qwen_generation_utils.py:418
Methodsync_handler
处理同步信号并更新任务状态
rft/trainer/zmq.py:478
Methodtokenize
Converts a string in a sequence of tokens. Args: text (`str`): The sequence to be encoded.
eval/utils/utils_odyssey/tokenization_qwen.py:253
Functiontop_k_logits
This function has been mostly taken from huggingface conversational ai code at https://medium.com/huggingface/how-to-build-a-state-of-th
eval/utils/utils_odyssey/qwen_generation_utils.py:388
Methodtraining_step
Perform a training step on a batch of inputs. Subclass and override to inject custom behavior. Args: model (`nn
sft/trainer.py:174
Methodvocab_size
(self)
eval/utils/utils_odyssey/tokenization_qwen.py:323
Methodwork_stealing
从其他节点窃取任务
rft/trainer/zmq.py:435
← previous501–568 of 568, ranked by callers